PROGRESS ON RFQIII FABRICATION IN J-PARC LINAC

T. Morishita, Y. Kondo, K. Hasegawa, JAEA, Tokai, Japan H. Kawamata, T. Sugimura, F. Naito, KEK, Tsukuba, Japan

Abstract For the beam current upgrade in J-PARC linac, the

fabrication of a new RFQ, which is designed for 50mA beam acceleration, has been started. The engineering design and the fabrication technologies were carefully chosen to reduce the discharge risk during the operation. For good vacuum pumping, vanes and ports are brazed for the direct pumping through slits at the tuners. Also, we tried a chemical polishing to improve the smoothness of the vane surface.

INTRODUCTION The J-PARC accelerator comprises an injector linac, a

3-GeV Rapid-Cycling Synchrotron and a 50-GeV Main Ring. The J-PARC linear accelerator consists of an ion source, an RFQ, DTLs, separated DTLs (SDTL), and the beam transport line to the RCS synchrotron [1].

The J-PARC linac has been operating for users with the beam energy of 181 MeV. The currently operating RFQ is a four-vane type cavity used to accelerate a negative hydrogen beam from 50 keV to 3 MeV with peak current of 30mA. The RF duty factor is 3% (600 μs at50 Hz). For the quick replacement in case of the RFQ trouble, we fabricated a spare RFQ (RFQII) as a backup machine. [2, 3, 4]

The energy (to 400MeV) and current (to 50mA) upgrade of the linac is scheduled for 1MW operation at RCS. The beam dynamics of an RFQ is newly designed for beam current upgrade to 50 mA operation. [4, 5] Then, we started a fabrication of RFQIII last year.

In this paper, we present the fabrication progress of an RFQIII for J-PARC linac.

ENGENEERING DESIGN Table 1 shows the RFQ III parameters. For the beam

current upgrade, the vane length is longer than RFQII about 0.5 m. The RFQ cavity is divided into three unit tanks. Each tank (about 1.2 m long) will be integrated together on a platform after the brazing of the major and the minor vanes.

Engineering design topics are listed in Table 2. The high-power test of the RFQII has been done successfully at April 2012. Basically the mechanical design and the fabrication procedure are the same as RFQII. Followings are the points which are changed from RFQII

Drilled Hole Plugging The drilled hole plugging technique was changed from

electron beam welding to the brazing. In the fabrication of RFQII, there was a vacuum leak at the welding spot after

the brazing (one of seventy-two welding points). Then we changed it to brazing along with the vane brazing in the RFQIII fabrication.

Table 1: Main Parameters of RFQIII

parameter RFQ III

Beam current [mA] 50

Frequency [MHz] 324

Acceleration energy [MeV] 0.05 to 3

Vane length [m] 3.6

Inter-vane voltage[kV] 81

Maximum surface field [MV/m] 30.7 (1.72 Kilpatrick)

Average bore radius [mm] 3.5

Vane-tip curvature [mm] 0.75r0(2.617)

Table 2: Mechanical Design Features

Material High-purity oxygen-free copper

with HIP(Hot Isostatic Pressing)

Drilled hole plugging

RFQII : Electron beam welding

RFQIII : Brazing

Annealing 600 degree C in vacuum furnace

Vane machining Numerical-controlled machining with ball-end mill

RFQIII: Introduced non-contact measurement [6]

Surface treatment Chemical polishing (3-5μm)

Integration method Vanes and ports are jointed in one step brazing

Unit cavities connection

RFQII : Welding for vacuum sealing, bolting for mechanical alignment

RFQIII : bolting for mechanical alignment and vacuum sealing

Dry Cutting for the Fitting of End Flanges Figure 1 shows a schematic drawing of the unit cavity

components to be assembled for brazing. After the final cutting of vanes, those are assembled and machined for a flange fitting at each ends. In the RFQII, vanes were disassembled for the surface treatment to remove cutting oil, then, reassembled for brazing. In the RFQIII, to reduce costs and production period, dry-machining for a flange fitting is adopted after the surface treatment. Then we can reduce the disassemble-reassemble process. ___________________________________________

#takatoshi.morishita@j-parc.jp

TUPB042 Proceedings of LINAC2012, Tel-Aviv, Israel

ISBN 978-3-95450-122-9

570Cop

yrig

htc ○

2012

byth

ere

spec

tive

auth

ors—

ccC

reat

ive

Com

mon

sAtt

ribu

tion

3.0

(CC

BY

3.0)

02 Proton and Ion Accelerators and Applications

2B Ion Linac Projects

Additionally, the displacements by the reassembling could be removed.

Unit Cavities Connection Unit cavities are bolted together with a vacuum seal

and an RF conductor. This configuration is simple and decomposable easily. However, the angle error between flanges at the connection will cause the alignment errors for the straightness between unit tanks. In the experience from RFQII fabrication, the angle error of the brazed flange was less than 0.07mm/(flange diameter). Then we put the 0.1 mm gap between flanges for a margin to correct the angle error. By adjusting the gap locally within the squashing tolerance of a vacuum seal and an RF contactor, t he angle error will be compensated.

Dividing Point of the Unit Cavity The longitudinal dividing point (i.e. the edges of the

second unit cavity) of the RFQII cavity is taken at the top/bottom of the modulation. In the RFQIII, dividing point is taken at the position where the longitudinal effect for the beam bunch is small. The gaps between unit cavities are 0.3 mm.

SURFACE TREATMENTS To reduce the discharge risk in the high electric field,

the inner surface of the cavity was chemically polished after the machining. The surface treatment procedure is summarized in Table 3. Figure 2 shows a chemical-polishing major vane in the bath.

Four kinds of solutions were tested before the fabrication of RFQII. Figure 3 shows the surface conditions for solutions and polishing depths.

• Solution #3 cannot produce a smooth surface. • A smooth surface can be obtained by the solution #4,

however, outshoots are found occasionally.

• Grain-boundary appears for over-polished (more than 10 μm) with the solutions #1,2, and 3.

• Lightly-polished (less than 10 μm) surface seems well with the solutions #1 and 4.

Table 3: Surface Treatment Procedure

Process Tools/Solutions procedure

Cleaning Water, methyl alcohol, MEK

wiping

Masking Taping, caulking Screw holes, brazing surface

Degreasing Alkaline solution Submerging

Acid wash Sulfuric acid aqueous solution

Submerging

Chemical polishing

Nitric acid + Phosphoric acid + Surfactant

Polishing depth is controlled by temperature and submerging time

Acid wash Sulfuric acid aqueous solution

Submerging

Anticorrosion Chromate solution Submerging

Cleaning /drying-out

Ultrapure water / drie

Running water

Air blow

Figure 4 shows the surface roughness for the cases of

before polishing, 3μm polished, and 10μm polished by the solution #4.

Figure 1: Assembling of vanes and other components of the unit tank for brazing.

Figure 2: Major vane is submerging in solution #1.

Major vane

Minor vane

Ports for cooling water

Ports for tuners/coupler

Ports for monitors

Flanges for endplate /tank connection

Flange fitting

Proceedings of LINAC2012, Tel-Aviv, Israel TUPB042

02 Proton and Ion Accelerators and Applications

2B Ion Linac Projects

ISBN 978-3-95450-122-9

571 Cop

yrig

htc ○

2012

byth

ere

spec

tive

auth

ors—

ccC

reat

ive

Com

mon

sAtt

ribu

tion

3.0

(CC

BY

3.0)

• Calculated average roughness (Ra) improved by the light polishing. The difference between 3μm and 10μm is small.

• Maximum height (Ry) decreased from 1.8μm to 1.2μm by 3μm polishing. Another 7μm polishing, the effect is 0.28μm.

Based on these examinations, we chose the solution #1 and 5μm polishing for the surface treatment.

PROGRESS OF RFQIII FABRICATION Final machining and surface treatment for first and

second units have been finished. The brazing of the first unit is in process. The machining of vanes for third unit is in the finishing step.

SCHEDULE The third unit will be brazed this November. After the

assembling, the RF tuning will be done at end of this year. The high-power test is scheduled next spring. The RFQIII will be installed in the accelerator tunnel at the shutdown period in 2013 after the off-line beam test with ion source and diagnostic beam line on the ground.

ACKNOWLEDGMENT The authors would like to thank Mr. Y. Iino and Mr. H.

Hatanaka of TOYAMA Co., Ltd. for the mechanical design and fabrication of RFQs II and III. It should also be acknowledged that Mr. T. Yoshida of Asahi Kinzoku Kogyo Inc., Y. Hori and Y. Saito of KEK to establish a method of the surface treatment process.

REFERENCES [1] Y. Yamazaki (eds), “Accelerator Technical Design

Report for J-PARC”JAERI-Tech 2003-044; KEK-Report 2002-13.

[2] T. Morishita et al., “Fabrication of the New RFQ for the J-PARC Linac”, Proc. of 2010 International Particle Accelerator Conference, Kyoto, Japan, p. 783 (2010).

[3] T. Morishita et al., Procs. of LINAC2010, Tsukuba, Japan, p. 518 and p.521 (2010).

[4] Y. Kondo et al., THPB052, this conference. [5] Y. Kondo et al., Phys. review STAB, to be published. [6] Y. Kondo et al., NIMA, 667, p. 5(2011).

Figure 3: Close-up view of polished surface. Test pieces are machined with the same condition of the vane.

Figure 4: Surface roughness before/after the polishing. Test pieces are machined with the same condition with the vane tip.

TUPB042 Proceedings of LINAC2012, Tel-Aviv, Israel

ISBN 978-3-95450-122-9

572Cop

yrig

htc ○

2012

byth

ere

spec

tive

auth

ors—

ccC

reat

ive

Com

mon

sAtt

ribu

tion

3.0

(CC

BY

3.0)

02 Proton and Ion Accelerators and Applications

2B Ion Linac Projects